/*-----------------------------------------------------------------------------*
DO YOU HAVE COVID19? HOW TO INCREASE THE USE OF DIAGNOSTIC AND CONTACT TRACING APPS
PURPOSE: Cleaning Dataset
Corresponding author: Carlos Scartascini (carlossc@iadb.org)
Last Update: May 17, 2021
Software Version: Stata 16 
*-----------------------------------------------------------------------------*/

drop _all
clear all
clear mata
clear matrix
set more off

/*
** INSTRUCTIONS
1. Unzip the data folder in your PC
2. Complete the information in the "DEFINE DIRECTORIES" section, be aware that you need to know your username to do so. 
3. Install the following packages if your don't have them already

*ssc install texdoc, replace
*ssc install outreg2, replace
*net install grc1leg.pkg

4. Run the code
*/

*---------------------------------------*
* 0. DEFINE DIRECTORIES					*
*---------------------------------------*

if "`c(username)'"=="SUSANAOT" {
	di in red "Susana"
	glo dir 	= "C:\Users\SUSANAOT\OneDrive - Inter-American Development Bank Group\SUSANAOT\99_Other\PLOS_CovidApps"
	glo data	= "$dir\data_cleaned"
	glo results	= "$dir\results"
}



*---------------------------------------*
* 1. OPEN DATASET AND DECLARE ITS TYPE	*
*---------------------------------------*
cd "$data"
use "COVID_APPS_Mexico.dta", clear 

*---------------------------------------*
* 2. GLOBALS 							*
*---------------------------------------*

** Dependent variables
glo dep_dis 	"app_alertas_dis app_sintomas_dis"
glo dep_dum 	"app_alertas_dum app_sintomas_dum"

** Mechanism 
glo mechanism 		"fb_prompted tramites_prompted seguridad_prompted"
glo mechanism_comp	"fb_prompted_comp tramites_prompted_comp seguridad_prompted_comp"

** Control variables (balance and regressions)
glo controls_bal 	"age young older female education college expositionCovid deathCovid older65 pr_contagio pr_hospital Party Visit risk_inside SocDist"
glo controls 		"age female education  expositionCovid deathCovid older65 pr_contagio pr_hospital Party Visit risk_inside SocDist"

** Heterogeneity 
glo heterog_d 	"Party Visit Risk_Inside SocDis"

** Treatment variables 
glo T "T_fb T_pro T_sec"


*---------------------------------------*
* MAIN RESULTS 							*
*---------------------------------------*
cd "$results"

** Table 1. Balance Table 
if 1==1{
cap texdoc close
	texdoc init Table1_Balance, replace force
	tex \begin{table}[H]
	tex \centering
	tex \scriptsize		
	tex \caption{Balance Table\label{tab:Balance}}
	tex \begin{tabular}{l*{5}{c}}			
	tex \toprule
	tex 					& Control & \multicolumn{3}{c}{Difference w.r.t. control} & Observations \\
	tex \cmidrule(lr){3-5}
	tex \textbf{Variable}  	& (av. \& s.e.) & T1 	& T2 	& T3 & \\
	tex & (1) & (2) & (3) & (4) & (5) \\
	tex \midrule
	foreach v in $controls_bal{
	disp in red "`v'"
		
		reg `v' $T, robust
		loc meT0`v'	: di %7.3f _b[_cons]
		loc seT0`v'	: di %7.3f _se[_cons]
		loc diT1`v' : di %7.3f _b[T_fb]
		loc seT1`v' : di %7.3f _se[T_fb]
		loc diT2`v' : di %7.3f _b[T_pro]
		loc seT2`v' : di %7.3f _se[T_pro]
		loc diT3`v' : di %7.3f _b[T_sec]
		loc seT3`v' : di %7.3f _se[T_sec]
		
		loc tbef1 = _b[T_fb]/_se[T_fb]
		loc pbef1 : di %7.3f 2*ttail(e(df_r),abs(`tbef1'))	
		loc staru1 = ""
		if ((`pbef1' < 0.1))  loc staru1 = "*" 
		if ((`pbef1' < 0.05)) loc staru1 = "**" 
		if ((`pbef1' < 0.01)) loc staru1 = "***" 
		
		loc tbef2 = _b[T_pro]/_se[T_pro]
		loc pbef2 : di %7.3f 2*ttail(e(df_r),abs(`tbef2'))	
		loc staru2 = ""
		if ((`pbef2' < 0.1))  loc staru2 = "*" 
		if ((`pbef2' < 0.05)) loc staru2 = "**" 
		if ((`pbef2' < 0.01)) loc staru2 = "***" 
			
		loc tbef3 = _b[T_sec]/_se[T_sec]
		loc pbef3 : di %7.3f 3*ttail(e(df_r),abs(`tbef3'))	
		loc staru3 = ""
		if ((`pbef3' < 0.1))  loc staru3 = "*" 
		if ((`pbef3' < 0.05)) loc staru3 = "**" 
		if ((`pbef3' < 0.01)) loc staru3 = "***" 
			
		loc n`v' : di %7.0f e(N)
	
		tex \parbox[l]{3.5cm}{ `:variable label `v''} 	& `meT0`v'' 		& `diT1`v''`staru1' 	& `diT2`v''`staru2' & `diT3`v''`staru3' & `n`v''\\
		tex                           				  	&  (`seT0`v'') 		& (`seT1`v'')	& (`seT2`v'') 	& (`seT3`v'')  & 	  \\
	}	
	tex \addlinespace[2pt] 
	tex \bottomrule
	tex \addlinespace[2pt]
	tex \multicolumn{6}{c}{\footnotesize{\begin{minipage}{0.85\textwidth}\textit{Notes:} Each row shows statistics for a different observable variable we have. Survey questions that serve the basis for the variables here, are available in Appendix C. Column [1] shows the sample average and the standard deviation in parentheses for the control group. Columns [2]-[4] show the regression coefficient and the standard error in parentheses corresponding to an OLS regression. Column [5] shows the sample size for each regression. Standard errors are robust. *** p<0.01, ** p<0.05, * p<0.1 \textit{Source:} Authors'calculations. \end{minipage}}}
	tex \end{tabular}
	tex \end{table}
	texdoc close	
}


** Table 2. Willingness to download the app
if 1==1{

foreach outcome in $dep_dum{

	reg `outcome' $T, robust
	test T_fb=T_pro=T_sec
	loc t1t2t3 : di %7.3f r(p)
	
	test T_fb=T_pro
	loc t1t2 : di %7.3f r(p)
	
	test T_fb=T_sec
	loc t1t3 : di %7.3f r(p)
	
	test T_pro=T_sec
	loc t2t3 : di %7.3f r(p)
	
	outreg2 using "Table2_ATE.tex", append keep($T) addstat(T1=T2=T3, `t1t2t3', T1=T2 ,`t1t2',T1=T3,`t1t3',T2=T3,`t2t3') addtext(Controls, No, Fixed Effects, No) dec(3) label addnote("\begin{minipage}{\textwidth}\textit{Notes:} Each row shows the regression coefficients and the standard error in parenthesis corresponding to an OLS regression. Dependent variables take the value 0-1 according to the willingness of the respondent to download each application. Survey questions used for the construction of the dependent variables are available in Appendix C. Standard errors are robust. *** p<0.01, ** p<0.05, * p<0.1. Controls include: sex, age, education, exposed to Covid, death to Covid, older than 65 at home, knows infected H1N1, belief about infection probability, belief about hospitalization probability, attends party, visits family, risk inside evaluation, and others practice social distancing. Survey questions used for the construction of the control variables available in Appendix C. \textit{Source:} Authors'calculations. \end{minipage}")

	
	reg `outcome' $T $controls, robust
	test T_fb=T_pro=T_sec
	loc t1t2t3 : di %7.3f r(p)
	
	test T_fb=T_pro
	loc t1t2 : di %7.3f r(p)
	
	test T_fb=T_sec
	loc t1t3 : di %7.3f r(p)
	
	test T_pro=T_sec
	loc t2t3 : di %7.3f r(p)
	
	outreg2 using "Table2_ATE.tex", append keep($T) addstat(T1=T2=T3, `t1t2t3', T1=T2 ,`t1t2',T1=T3,`t1t3',T2=T3,`t2t3') addtext(Controls, Yes, Fixed Effects, No) dec(3) label 
	
	
	reg `outcome' $T $controls i.state, robust
	test T_fb=T_pro=T_sec
	loc t1t2t3 : di %7.3f r(p)
	
	test T_fb=T_pro
	loc t1t2 : di %7.3f r(p)
	
	test T_fb=T_sec
	loc t1t3 : di %7.3f r(p)
	
	test T_pro=T_sec
	loc t2t3 : di %7.3f r(p)
	
	outreg2 using "Table2_ATE.tex", append keep($T) addstat(T1=T2=T3, `t1t2t3', T1=T2 ,`t1t2',T1=T3,`t1t3',T2=T3,`t2t3') addtext(Controls, Yes, Fixed Effects, State) dec(3) label 


	reg `outcome' $T $controls i.muni, robust
	test T_fb=T_pro=T_sec
	loc t1t2t3 : di %7.3f r(p)
	
	test T_fb=T_pro
	loc t1t2 : di %7.3f r(p)
	
	test T_fb=T_sec
	loc t1t3 : di %7.3f r(p)
	
	test T_pro=T_sec
	loc t2t3 : di %7.3f r(p)
	
	outreg2 using "Table2_ATE.tex", append keep($T) addstat(T1=T2=T3, `t1t2t3', T1=T2 ,`t1t2',T1=T3,`t1t3',T2=T3,`t2t3') addtext(Controls, Yes, Fixed Effects, Muni) dec(3) label 
}

}


** Figure 1. Treatment Effects and Coeffs estimates
if 1==1{
	eststo Tracing: reg app_alertas_dum $T $controls, robust
	eststo Diagnostic: reg  app_sintomas_dum $T $controls, robust

	coefplot Tracing || Diagnostic, keep(T_fb T_pro T_sec age female education expositionCovid deathCovid older65 h1n1 pr_contagio pr_hospital Party Visit risk_inside SocDist) relocate(T_fb=1 T_pro=2 T_sec=3  age=5 female=6 education=7 expositionCovid=8 deathCovid=9 older65=10 pr_contagio=11 pr_hospital=12 Party=13 Visit=14 risk_inside=15 SocDist=16) levels(95) xline(0)  drop(_cons) yline(0) ciopts(recast(rcap)) citop  note("Note: 95% confidence interval")
	graph export Fig1.pdf, as(pdf) replace
	graph export Fig1.eps, as(eps) replace
	graph export Fig1.tif, as(tif) replace
}


** Figure 2. Treatment Effects - Ordered Logit
if 1==1{
	forval n=1(1)4{
		ologit app_alertas_dis $T , robust
		margins, dydx($T) predict(outcome(`n')) level(95) post
		estimates store T`n'
	}
	
	coefplot (T1, label(Definitely not) msymbol(X)) (T2, label(Don't think so) msymbol(D) ) (T3, label(Think so) msymbol(Sh)) (T4, label(Definitely yes) ) , levels(95) vertical drop(_cons) yline(0) ciopts(recast(rcap)) citop ytitle("Marginal effect [Contact Tracing]") note("Note: 95% confidence interval") name(Fig2A, replace)
	
	forval n=1(1)4{
		ologit app_sintomas_dis $T , robust
		margins, dydx($T) predict(outcome(`n')) level(95) post
		estimates store D`n'
	}

	coefplot (D1, label(Definitely not) msymbol(X)) (D2, label(Don't think so) msymbol(D) ) (D3, label(Think so) msymbol(Sh)) (D4, label(Definitely yes) )  , levels(95) vertical drop(_cons) yline(0) ciopts(recast(rcap)) citop ytitle("Marginal effect [Diagnostic]") note("Note: 95% confidence interval") name(Fig2B, replace)
	
	grc1leg Fig2A Fig2B, legendfrom(Fig2A) c(1) ycom name(Fig2, replace)
	graph display Fig2, xsize(15) ysize(20)
	graph export Fig2.pdf, as(pdf) replace
	graph export Fig2.eps, as(eps) replace
	graph export Fig2.tif, as(tif) replace
}


** Figure 3. Treatment Effects - Sonora sample
if 1==1{
	preserve 
		use "$data\COVID_APPS_Sonora.dta", clear 

		eststo Tracing: reg app_alertas_dum T_fb T_pro T_sec_o T_sec_n, robust
		eststo Diagnostic: reg  app_sintomas_dum T_fb T_pro T_sec_o T_sec_n, robust

		coefplot Tracing || Diagnostic, keep(T_fb T_pro T_sec_o T_sec_n) relocate(T_fb=1 T_pro=2 T_sec_o=3 T_sec_n=4) levels(90) xline(0)  drop(_cons) yline(0) ciopts(recast(rcap)) citop  note("Note: 90% confidence interval")
		graph export Fig3.pdf, as(pdf) replace	
		graph export Fig3.eps, as(eps) replace
		graph export Fig3.tif, as(tif) replace	
	restore
}


** Figure A3. Distribution of responses - control group
if 1==1{
	twoway (hist app_alertas_dis if T0==1, discrete frac lcolor(gs12) fcolor(gs12))(hist app_sintomas_dis if T0==1, discrete frac fcolor(none) lcolor(red)), title("Contact Tracing App (grey mass) vs Diagnostic App (red line)", size(small)) ytitle(, size(small)) ylabel(, labsize(vsmall)) legend(off) xlabel(1(1)4,labsize(vsmall)) xtitle("1 less likely - 4 more likely" , size(small))
	graph export S3_Fig.pdf, as(pdf) replace	
}


** Table B1. Willingness to download the app [Ordered Logistic Regression]
if 1==1{
	foreach y in $dep_dis {

		ologit `y' $T , robust 
		test T_fb=T_pro=T_sec
		loc t1t2t3 : di %7.3f r(p)
		
		test T_fb=T_pro
		loc t1t2 : di %7.3f r(p)
		
		test T_fb=T_sec
		loc t1t3 : di %7.3f r(p)
		
		test T_pro=T_sec
		loc t2t3 : di %7.3f r(p)

		outreg2 using "TableB1_OLogit.tex", append keep($T) addstat(T1=T2=T3, `t1t2t3', T1=T2 ,`t1t2',T1=T3,`t1t3',T2=T3,`t2t3') addtext(Controls, No, Fixed Effects, No) dec(3) label addnote("\begin{minipage}{\textwidth}\textit{Notes:} Each row shows the regression coefficients and the standard error in parenthesis corresponding to an ordered logit regression. Dependent variables take the values 1 (definitely would not) to 4 (definitely would) according to the willingness of the respondent to download each one of the apps. Survey questions used for the construction of the dependent variables available in Appendix C. Standard errors are robust. *** p<0.01, ** p<0.05, * p<0.1. Controls include: sex, age, education, exposed to Covid, death to Covid, older than 65 at home, knows infected H1N1, belief about infection probability, belief about hospitalization probability, attends party, visits family, risk inside evaluation, and others practice social distancing. Survey questions used for the construction of the control variables available in Appendix C. \end{minipage}")


		ologit `y' $T $controls, robust 
		test T_fb=T_pro=T_sec
		loc t1t2t3 : di %7.3f r(p)
		
		test T_fb=T_pro
		loc t1t2 : di %7.3f r(p)
		
		test T_fb=T_sec
		loc t1t3 : di %7.3f r(p)
		
		test T_pro=T_sec
		loc t2t3 : di %7.3f r(p)

		outreg2 using "TableB1_OLogit.tex", append keep($T) addstat(T1=T2=T3, `t1t2t3', T1=T2 ,`t1t2',T1=T3,`t1t3',T2=T3,`t2t3') addtext(Controls, Yes, Fixed Effects, No) dec(3) label 

		
		ologit `y' $T $controls i.state, robust 
		test T_fb=T_pro=T_sec
		loc t1t2t3 : di %7.3f r(p)
		
		test T_fb=T_pro
		loc t1t2 : di %7.3f r(p)
		
		test T_fb=T_sec
		loc t1t3 : di %7.3f r(p)
		
		test T_pro=T_sec
		loc t2t3 : di %7.3f r(p)
		
		outreg2 using "TableB1_OLogit.tex", append keep($T) addstat(T1=T2=T3, `t1t2t3', T1=T2 ,`t1t2',T1=T3,`t1t3',T2=T3,`t2t3') addtext(Controls, Yes, Fixed Effects, State) dec(3) label 


	}
}


** Table B2. Balance - Sonora Sample 
cd "$data"
use "COVID_APPS_Sonora.dta", clear 

glo T_Sonora "T_fb T_pro T_sec_o T_sec_n"
glo controls "age female education expositionCovid deathCovid older65 pr_contagio pr_hospital Party Visit risk_inside SocDist"

cd "$results"
if 1==1{
cap texdoc close
	texdoc init TableB2_BalanceSonora, replace force
	tex \begin{table}[H]
	tex \centering
	tex \scriptsize		
	tex \caption{Balance Table\label{tab:Balance}}
	tex \begin{tabular}{l*{6}{c}}			
	tex \toprule
	tex 					& Control & \multicolumn{4}{c}{Difference w.r.t. control} & Observations \\
	tex \cmidrule(lr){3-5}
	tex \textbf{Variable}  	& (av. \& s.e.) & T1 	& T2 	& T3 & T4& \\
	tex & (1) & (2) & (3) & (4) & (5) & (6) \\
	tex \midrule
	foreach v in $controls_bal{
	disp in red "`v'"
		
		reg `v' $T_Sonora, robust
		loc meT0`v'	: di %7.3f _b[_cons]
		loc seT0`v'	: di %7.3f _se[_cons]
		loc diT1`v' : di %7.3f _b[T_fb]
		loc seT1`v' : di %7.3f _se[T_fb]
		loc diT2`v' : di %7.3f _b[T_pro]
		loc seT2`v' : di %7.3f _se[T_pro]
		loc diT3`v' : di %7.3f _b[T_sec_o]
		loc seT3`v' : di %7.3f _se[T_sec_o]
		loc diT4`v' : di %7.3f _b[T_sec_n]
		loc seT4`v' : di %7.3f _se[T_sec_n]
		
		loc tbef1 = _b[T_fb]/_se[T_fb]
		loc pbef1 : di %7.3f 2*ttail(e(df_r),abs(`tbef1'))	
		loc staru1 = ""
		if ((`pbef1' < 0.1))  loc staru1 = "*" 
		if ((`pbef1' < 0.05)) loc staru1 = "**" 
		if ((`pbef1' < 0.01)) loc staru1 = "***" 
		
		loc tbef2 = _b[T_pro]/_se[T_pro]
		loc pbef2 : di %7.3f 2*ttail(e(df_r),abs(`tbef2'))	
		loc staru2 = ""
		if ((`pbef2' < 0.1))  loc staru2 = "*" 
		if ((`pbef2' < 0.05)) loc staru2 = "**" 
		if ((`pbef2' < 0.01)) loc staru2 = "***" 
			
		loc tbef3 = _b[T_sec_o]/_se[T_sec_o]
		loc pbef3 : di %7.3f 3*ttail(e(df_r),abs(`tbef3'))	
		loc staru3 = ""
		if ((`pbef3' < 0.1))  loc staru3 = "*" 
		if ((`pbef3' < 0.05)) loc staru3 = "**" 
		if ((`pbef3' < 0.01)) loc staru3 = "***" 
		
		loc tbef4 = _b[T_sec_n]/_se[T_sec_n]
		loc pbef4 : di %7.4f 4*ttail(e(df_r),abs(`tbef4'))	
		loc staru4 = ""
		if ((`pbef4' < 0.1))  loc staru4 = "*" 
		if ((`pbef4' < 0.05)) loc staru4 = "**" 
		if ((`pbef4' < 0.01)) loc staru4 = "***" 
			
		loc n`v' : di %7.0f e(N)
	
		tex \parbox[l]{3.5cm}{ `:variable label `v''} 	& `meT0`v'' 		& `diT1`v''`staru1' 	& `diT2`v''`staru2' & `diT3`v''`staru3' & `diT4`v''`staru4' & `n`v''\\
		tex                           				  	&  (`seT0`v'') 		& (`seT1`v'')	& (`seT2`v'') 	& (`seT3`v'') & (`seT4`v'') & 	  \\
	}	
	tex \addlinespace[2pt] 
	tex \bottomrule
	tex \addlinespace[2pt]
	tex \multicolumn{7}{c}{\footnotesize{\begin{minipage}{0.85\textwidth}\textit{Notes:} Each row shows statistics for a different observable variable we have. Survey questions that serve the basis for the variables here, are available in Appendix C. Column [1] shows the sample average and the standard deviation in parentheses for the control group. Columns [2]-[4] show the regression coefficient and the standard error in parentheses corresponding to an OLS regression. Column [5] shows the sample size for each regression. Standard errors are robust. *** p<0.01, ** p<0.05, * p<0.1 \textit{Source:} Authors'calculations. \end{minipage}}}
	tex \end{tabular}
	tex \end{table}
	texdoc close	
}


** Table B3. Willingness to download the app - Sonora Sample 
if 1==1{

foreach outcome in $dep_dum{

	reg `outcome' $T_Sonora, robust
	test T_fb=T_pro=T_sec_n=T_sec_o
	loc t1t2t3t4 : di %7.3f r(p)

	test T_fb=T_pro
	loc t1t2 : di %7.3f r(p)
	
	test T_fb=T_sec_o
	loc t1t3 : di %7.3f r(p)
	
	test T_fb=T_sec_n
	loc t1t4 : di %7.3f r(p)
	
	test T_pro=T_sec_o
	loc t2t3 : di %7.3f r(p)
	
	test T_pro=T_sec_n
	loc t2t4 : di %7.3f r(p)
	
	test T_sec_o=T_sec_n
	loc t3t4 : di %7.3f r(p)
	
	outreg2 using "TableB3_ATESonora.tex", append keep($T_Sonora) addstat(T1=T2=T3=T4, `t1t2t3t4', T1=T2 ,`t1t2',T1=T3,`t1t3', T1=T4, `t1t4',T2=T3,`t2t3', T2=T4, `t2t4', T3=T4, `t3t4') addtext(Controls, No, Fixed Effects, No) dec(3) label addnote("\begin{minipage}{\textwidth}\textit{Notes:} Each row shows the regression coefficients and the standard error in parenthesis corresponding to an OLS regression. Dependent variables take the value 0-1 according to the willingness of the respondent to download each application. Survey questions used for the construction of the dependent variables are available in Appendix C. Standard errors are robust. *** p<0.01, ** p<0.05, * p<0.1. Controls include: sex, age, education, exposed to Covid, death to Covid, older than 65 at home, knows infected H1N1, belief about infection probability, belief about hospitalization probability, attends party, visits family, risk inside evaluation, and others practice social distancing. Survey questions used for the construction of the control variables available in Appendix C. \textit{Source:} Authors'calculations. \end{minipage}")

	
	reg `outcome' $T_Sonora $controls, robust
	test T_fb=T_pro=T_sec_n=T_sec_o
	loc t1t2t3t4 : di %7.3f r(p)
	
	test T_fb=T_pro
	loc t1t2 : di %7.3f r(p)
	
	test T_fb=T_sec_o
	loc t1t3 : di %7.3f r(p)
	
	test T_fb=T_sec_n
	loc t1t4 : di %7.3f r(p)
	
	test T_pro=T_sec_o
	loc t2t3 : di %7.3f r(p)
	
	test T_pro=T_sec_n
	loc t2t4 : di %7.3f r(p)
	
	test T_sec_o=T_sec_n
	loc t3t4 : di %7.3f r(p)
	
	outreg2 using "TableB3_ATESonora.tex", append keep($T_Sonora) addstat(T1=T2=T3=T4, `t1t2t3t4', T1=T2 ,`t1t2',T1=T3,`t1t3', T1=T4, `t1t4',T2=T3,`t2t3', T2=T4, `t2t4', T3=T4, `t3t4') addtext(Controls, Yes, Fixed Effects, No) dec(3) label


	reg `outcome' $T_Sonora $controls i.muni, robust
	test T_fb=T_pro=T_sec_n=T_sec_o
	loc t1t2t3t4 : di %7.3f r(p)
	
	test T_fb=T_pro
	loc t1t2 : di %7.3f r(p)
	
	test T_fb=T_sec_o
	loc t1t3 : di %7.3f r(p)
	
	test T_fb=T_sec_n
	loc t1t4 : di %7.3f r(p)
	
	test T_pro=T_sec_o
	loc t2t3 : di %7.3f r(p)
	
	test T_pro=T_sec_n
	loc t2t4 : di %7.3f r(p)
	
	test T_sec_o=T_sec_n
	loc t3t4 : di %7.3f r(p)
	
	outreg2 using "TableB3_ATESonora.tex", append keep($T_Sonora) addstat(T1=T2=T3=T4, `t1t2t3t4', T1=T2 ,`t1t2',T1=T3,`t1t3', T1=T4, `t1t4',T2=T3,`t2t3', T2=T4, `t2t4', T3=T4, `t3t4') addtext(Controls, Yes, Fixed Effects, Muni) dec(3) label
}

}


